kreuzberg 4.0.0.pre.rc.16 → 4.0.0.pre.rc.17

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: fe3add89c26722e26baf090f7b9a0c32671c449be6a34ea4285a5f6d15548b72
-  data.tar.gz: 49147ceab3fddc3161ff0df55f7c535134d63da7ce2577aad905c91179e875f3
+  metadata.gz: f91977b1472bb6211f3ac2efad274e8cbc77dc5ed9832529eccbebeae1f74b4f
+  data.tar.gz: b8a32377a80cfec656e8ddd65576dc220f497fc65b7b45b307db54a0b3b4a274
 SHA512:
-  metadata.gz: 5f2e0ab3d3dd4c975a99dcbf4a2e81347673eb74687034f8ef72cc3ece6561fbbed70811edc7363c911385c2d7c2eb0be2d8fa990872845458a3d3f5f019422c
-  data.tar.gz: 530bf825eb92e9a3df838ab14ec68277b17e575833ecdf0af11e32d8749e101e4fe68195841524c5e6b41c31a2330076e9da4507f5edf047d8670ad26c9dd928
+  metadata.gz: cb391d9f82848e0b19b0c8df2cce7db455d1b73ba5e5c6dd63a2cc87732d8dd0cd6596ca7f9b305061d9400db95c5890292efbd16af7e55a9434f3f29a337642
+  data.tar.gz: 9e41afcc217e00d9feb3f8c4adecb7152743227f2c77f4bcfd9fd5e3d4b64b01171d3bdbb2b1290e10d2efdf13cdb53feb8fee01a95b6ba4d87ea76425b56692

data/Gemfile.lock

@@ -1,7 +1,7 @@
 PATH
   remote: .
   specs:
-    kreuzberg (4.0.0.pre.rc.16)
+    kreuzberg (4.0.0.pre.rc.17)
 
 GEM
   remote: https://rubygems.org/
@@ -33,7 +33,7 @@ GEM
     ffi (1.17.2-arm64-darwin)
     ffi (1.17.2-x86_64-linux-gnu)
     fileutils (1.8.0)
-    i18n (1.14.7)
+    i18n (1.14.8)
       concurrent-ruby (~> 1.0)
     json (2.18.0)
     language_server-protocol (3.17.0.5)

data/ext/kreuzberg_rb/native/Cargo.lock

@@ -2354,7 +2354,7 @@ dependencies = [
 
 [[package]]
 name = "kreuzberg-rb"
-version = "4.0.0-rc.16"
+version = "4.0.0-rc.17"
 dependencies = [
  "async-trait",
  "html-to-markdown-rs",

data/ext/kreuzberg_rb/native/Cargo.toml

@@ -7,7 +7,7 @@ rb-sys = { path = "../../../vendor/rb-sys" }
 
 [package]
 name = "kreuzberg-rb"
-version = "4.0.0-rc.16"
+version = "4.0.0-rc.17"
 edition = "2024"
 rust-version = "1.91"
 authors = ["Na'aman Hirschfeld <nhirschfeld@gmail.com>"]

data/ext/kreuzberg_rb/native/build.rs

@@ -1,52 +1,75 @@
-#[cfg(target_os = "macos")]
-fn main() {
-    if let Ok(cargo_manifest_dir) = std::env::var("CARGO_MANIFEST_DIR") {
-        let lib_path = std::path::Path::new(&cargo_manifest_dir)
-            .parent()
-            .and_then(|p| p.parent())
-            .and_then(|p| p.parent())
-            .and_then(|p| p.parent())
-            .and_then(|p| p.parent())
-            .map(|p| p.join("target/release"))
-            .expect("Failed to construct lib path");
-        println!("cargo:rustc-link-search={}", lib_path.display());
-    }
-    println!("cargo:rustc-link-arg=-Wl,-undefined,dynamic_lookup");
-    println!("cargo:rustc-link-arg=-Wl,-rpath,@loader_path");
-    println!("cargo:rustc-link-arg=-Wl,-rpath,@loader_path/.");
-}
+use std::env;
+use std::path::PathBuf;
 
-#[cfg(target_os = "linux")]
 fn main() {
-    if let Ok(cargo_manifest_dir) = std::env::var("CARGO_MANIFEST_DIR") {
-        let lib_path = std::path::Path::new(&cargo_manifest_dir)
-            .parent()
-            .and_then(|p| p.parent())
-            .and_then(|p| p.parent())
-            .and_then(|p| p.parent())
-            .and_then(|p| p.parent())
-            .map(|p| p.join("target/release"))
-            .expect("Failed to construct lib path");
-        println!("cargo:rustc-link-search={}", lib_path.display());
+    let target = env::var("TARGET").unwrap();
+    let profile = env::var("PROFILE").unwrap_or_else(|_| "release".to_string());
+
+    // Try to locate kreuzberg-ffi library built alongside this crate
+    let cargo_manifest_dir = env::var("CARGO_MANIFEST_DIR").unwrap();
+    let manifest_path = PathBuf::from(&cargo_manifest_dir);
+
+    // Prefer host target layout, but include target-triple layout for cross builds.
+    // IMPORTANT: Only search lib directories, NOT deps directories.
+    // The deps/ directories may contain dylibs with hardcoded install_name paths,
+    // which causes load errors on macOS when users install the gem.
+    if let Some(packages_root) = manifest_path
+        .parent()
+        .and_then(|p| p.parent())
+        .and_then(|p| p.parent())
+        .and_then(|p| p.parent())
+        .and_then(|p| p.parent())
+    {
+        let host_lib_dir = packages_root.join("target").join(&profile);
+        let target_lib_dir = packages_root.join("target").join(&target).join(&profile);
+
+        // Try to find the static library and link it directly on Unix-like systems
+        // to avoid the linker preferring dylib over static lib.
+        if !target.contains("windows") {
+            let static_lib_name = if target.contains("windows") {
+                "kreuzberg_ffi.lib"
+            } else {
+                "libkreuzberg_ffi.a"
+            };
+
+            // Check both host and target lib directories for the static library
+            for lib_dir in [&host_lib_dir, &target_lib_dir] {
+                let static_lib = lib_dir.join(static_lib_name);
+                if static_lib.exists() {
+                    // Found static library, link it directly by passing the full path
+                    println!("cargo:rustc-link-arg={}", static_lib.display());
+                    // Don't add the library search path or -l flag
+                    // Jump to platform-specific configuration
+                    if target.contains("darwin") {
+                        println!("cargo:rustc-link-arg=-Wl,-undefined,dynamic_lookup");
+                        println!("cargo:rustc-link-arg=-Wl,-rpath,@loader_path");
+                    } else if target.contains("linux") {
+                        println!("cargo:rustc-link-arg=-Wl,-rpath,$ORIGIN");
+                    }
+                    println!("cargo:rerun-if-changed=build.rs");
+                    return;
+                }
+            }
+        }
+
+        // Fallback: Add search paths and use standard linking
+        for dir in [host_lib_dir, target_lib_dir] {
+            if dir.exists() {
+                println!("cargo:rustc-link-search=native={}", dir.display());
+            }
+        }
     }
-    println!("cargo:rustc-link-arg=-Wl,-rpath,$ORIGIN");
-    println!("cargo:rustc-link-arg=-Wl,-rpath,$ORIGIN/.");
-}
 
-#[cfg(target_os = "windows")]
-fn main() {
-    if let Ok(cargo_manifest_dir) = std::env::var("CARGO_MANIFEST_DIR") {
-        let lib_path = std::path::Path::new(&cargo_manifest_dir)
-            .parent()
-            .and_then(|p| p.parent())
-            .and_then(|p| p.parent())
-            .and_then(|p| p.parent())
-            .and_then(|p| p.parent())
-            .map(|p| p.join("target/release"))
-            .expect("Failed to construct lib path");
-        println!("cargo:rustc-link-search={}", lib_path.display());
+    // Link the kreuzberg-ffi library
+    // When kreuzberg-ffi is built, its symbols become available for linking
+    println!("cargo:rustc-link-lib=static=kreuzberg_ffi");
+
+    if target.contains("darwin") {
+        println!("cargo:rustc-link-arg=-Wl,-undefined,dynamic_lookup");
+        println!("cargo:rustc-link-arg=-Wl,-rpath,@loader_path");
+    } else if target.contains("linux") {
+        println!("cargo:rustc-link-arg=-Wl,-rpath,$ORIGIN");
     }
-}
 
-#[cfg(not(any(target_os = "macos", target_os = "linux", target_os = "windows")))]
-fn main() {}
+    println!("cargo:rerun-if-changed=build.rs");
+}

data/ext/kreuzberg_rb/native/src/lib.rs

@@ -78,8 +78,7 @@ pub struct CMetadataField {
 }
 
 // These C ABI functions are provided by the kreuzberg-ffi crate
-// We declare them here to ensure proper linking on all platforms
-#[link(name = "kreuzberg_ffi", kind = "static")]
+// Linking is handled by build.rs to ensure static linking
 unsafe extern "C" {
     pub fn kreuzberg_last_error_code() -> i32;
     pub fn kreuzberg_last_panic_context() -> *mut c_char;
@@ -3071,7 +3070,7 @@ fn validate_chunking_params(max_chars: usize, max_overlap: usize) -> Result<i32,
 /// Gets valid binarization methods as a JSON string
 ///
 /// @return [String] JSON array of valid binarization methods
-fn get_valid_binarization_methods(ruby: &Ruby) -> Result<String, Error> {
+fn get_valid_binarization_methods(_ruby: &Ruby) -> Result<String, Error> {
     let ptr = unsafe { kreuzberg_get_valid_binarization_methods() };
     if ptr.is_null() {
         return Err(runtime_error("Failed to get valid binarization methods"));
@@ -3095,7 +3094,7 @@ fn get_valid_binarization_methods(ruby: &Ruby) -> Result<String, Error> {
 /// Gets valid language codes as a JSON string
 ///
 /// @return [String] JSON array of valid language codes
-fn get_valid_language_codes(ruby: &Ruby) -> Result<String, Error> {
+fn get_valid_language_codes(_ruby: &Ruby) -> Result<String, Error> {
     let ptr = unsafe { kreuzberg_get_valid_language_codes() };
     if ptr.is_null() {
         return Err(runtime_error("Failed to get valid language codes"));
@@ -3119,7 +3118,7 @@ fn get_valid_language_codes(ruby: &Ruby) -> Result<String, Error> {
 /// Gets valid OCR backends as a JSON string
 ///
 /// @return [String] JSON array of valid OCR backends
-fn get_valid_ocr_backends(ruby: &Ruby) -> Result<String, Error> {
+fn get_valid_ocr_backends(_ruby: &Ruby) -> Result<String, Error> {
     let ptr = unsafe { kreuzberg_get_valid_ocr_backends() };
     if ptr.is_null() {
         return Err(runtime_error("Failed to get valid OCR backends"));
@@ -3143,7 +3142,7 @@ fn get_valid_ocr_backends(ruby: &Ruby) -> Result<String, Error> {
 /// Gets valid token reduction levels as a JSON string
 ///
 /// @return [String] JSON array of valid token reduction levels
-fn get_valid_token_reduction_levels(ruby: &Ruby) -> Result<String, Error> {
+fn get_valid_token_reduction_levels(_ruby: &Ruby) -> Result<String, Error> {
     let ptr = unsafe { kreuzberg_get_valid_token_reduction_levels() };
     if ptr.is_null() {
         return Err(runtime_error("Failed to get valid token reduction levels"));
@@ -3395,7 +3394,7 @@ fn get_error_details_native(ruby: &Ruby) -> Result<Value, Error> {
     // SAFETY: FFI function is thread-safe and returns a struct with allocated C strings
     let details = unsafe { kreuzberg_get_error_details() };
 
-    let hash = RHash::new();
+    let hash = ruby.hash_new();
 
     // Convert C strings to Ruby strings, handling nulls safely
     // SAFETY: All non-null pointers from FFI must be valid C strings

data/lib/kreuzberg/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Kreuzberg
-  VERSION = '4.0.0-rc.16'
+  VERSION = '4.0.0-rc.17'
 end

data/vendor/Cargo.toml

@@ -2,7 +2,7 @@
 members = ["kreuzberg", "kreuzberg-tesseract"]
 
 [workspace.package]
-version = "4.0.0-rc.16"
+version = "4.0.0-rc.17"
 edition = "2024"
 rust-version = "1.91"
 authors = ["Na'aman Hirschfeld <nhirschfeld@gmail.com>"]

data/vendor/kreuzberg/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "kreuzberg"
-version = "4.0.0-rc.16"
+version = "4.0.0-rc.17"
 edition = "2024"
 rust-version = "1.91"
 authors = ["Na'aman Hirschfeld <nhirschfeld@gmail.com>"]

data/vendor/kreuzberg/src/extraction/html.rs

@@ -163,6 +163,7 @@ fn convert_html_with_options_large_stack(html: String, options: ConversionOption
 }
 
 #[cfg(not(target_arch = "wasm32"))]
+#[allow(dead_code)]
 fn convert_inline_images_with_large_stack(
     html: String,
     options: ConversionOptions,
@@ -204,20 +205,6 @@ fn extract_panic_reason(panic: &Box<dyn Any + Send + 'static>) -> String {
 }
 
 // WASM implementations skip dedicated stack (not supported) and process inline
-#[cfg(target_arch = "wasm32")]
-fn convert_html_with_options_large_stack(html: String, options: ConversionOptions) -> Result<String> {
-    convert_html_with_options(&html, options)
-}
-
-#[cfg(target_arch = "wasm32")]
-fn convert_inline_images_with_large_stack(
-    html: String,
-    options: ConversionOptions,
-    image_config: LibInlineImageConfig,
-) -> Result<HtmlExtraction> {
-    convert_inline_images_with_options(&html, options, image_config)
-}
-
 /// Convert HTML to markdown with optional configuration.
 ///
 /// Uses sensible defaults if no configuration is provided:

data/vendor/kreuzberg-tesseract/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "kreuzberg-tesseract"
-version = "4.0.0-rc.16"
+version = "4.0.0-rc.17"
 edition = "2024"
 rust-version = "1.91"
 authors = ["Na'aman Hirschfeld <nhirschfeld@gmail.com>"]

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: kreuzberg
 version: !ruby/object:Gem::Version
-  version: 4.0.0.pre.rc.16
+  version: 4.0.0.pre.rc.17
 platform: ruby
 authors:
 - Na'aman Hirschfeld
 autorequire:
 bindir: exe
 cert_chain: []
-date: 2025-12-21 00:00:00.000000000 Z
+date: 2025-12-22 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: bundler